J . Am. Chem. SOC.1987, 109, 380-386
380
An Unusual Copper-Uridine Octamer: Existence in Solution and Structural Study in Its Auto-Built Zeolitic Network in the Solid State J. Galy,*" A. Mosset,la I. Grenthe,*lbI. Puigdom6nech,+lbB. Sjoberg,lEand F. Hult6dd Contribution f r o m the Laboratorie de Chimie de Coordination du C N R S , F-31400 Toulouse, France, the Department of Inorganic Chemistry, Royal Institute of Technology, S - 100 44 Stockholm, Sweden, the Department of Medical Biochemistry, University of Goteborg, S-400 33 Coteborg, Sweden, and the Department of Inorganic Chemistry, Chalmers University of Technology, S-412 96 Goteborg, Sweden. Received April 28, 1986
Abstract: The structure of a copper(I1)-uridine complex has been determined from a single-crystal X-ray structure determination. The compound C~~(urid)~Na~-SNaC10~.48H~O (urid = uridine(3-)) crystallizes in a cubic cell of dimensions a = 23.716 A with 2 = 3 and space group P432. The complex forms an octamer of formula [ C ~ ~ ( u r i d ) ~ N a ( H ~ O )this ~ ] ' molecular -; species exhibits a toric structure. The uridine ligand is three-coordinated toward three copper atoms by N(3), 0 ( 3 ' ) , and 0(2'), this last oxygen atom bridging two copper atoms. The conformation about the glycosidic bond is anti, the puckering of the ribose ring is C(2') endo, and the conformation about C(4')-C(5') is gauche-gauche. The sodium atoms form a three-dimensional network quite similar to the framework of the synthetic Linde A zeolite. Two kinds of cavities are thus defined: chambers, occupied by water molecules and sodium cations, and tunnels, occupied by perchlorate groups and the octamer copper-uridine complex. In aqueous solutions, the existence of either the same octamer or a similar complex has been demonstrated by both potentiometric titrations and small-angle X-ray scattering measurements.
The biological roles of nucleic acids and nucleotides are dependent on metal ions; some examples are the use of certain transition-metal complexes (particularly platinum compounds) in the chemotherapy of some tumors and the involvement of metal cations in biological processes which distinguish between r i b - and deoxyribonucleotides.2 Many studies have been made of the interactions between metal ions, nucleosides, and nucleotides in order to understand the much more complicated nucleic acidmetal systems. The building blocks of nucleic acids contain a large number of oxygen and nitrogen donor sites of varying basicity and, hence, of varying donor properties. An increasing number of crystallographic studies are progressively bringing out information of the general bonding types of different metal ions in complexes of nucleic acid constituent^.^^^ There are also a number of studies made in solution, mainly by N M R technique and conventional p~tentiometry.~ Nevertheless, the information about nucleoside complexes is scarce, due to the lack of both precise structural and solution chemical data. This is in part due to the weak stability of the complexes f ~ r m e d which ,~ makes it difficult both to interpret solution chemical data and to obtain suitable single crystals. W e have selected the copper(I1)-uridine system in order to obtain data on the structure and bonding in a metal nucleoside system both in the solid state and in solution. Literature studies on the complex formation between Cu(I1) and uridine in aqueous solutions are scarce and contradictory. Fiskin and Beer6 determined the equilibrium constant for the 1:l complex by measuring proton displacement in acid (=pH 5 ) solutions. On the other hand, Reinert and Weiss7 found no complex formation a t < p H 6, but above =pH 8 they found a soluble 1:l green-blue complex. Chao and Kearns8 studied water-Me2S0 solutions of the Cu(I1)-uridine system with ESR spectroscopy, establishing the existence of a diamagnetic (or antiferromagnetic9) complex between p H 8.2 and 10. Berger and Eichhorn'O found, with the N M R technique, that uridine binds Cu2+ both a t N(3) and a t the ribose hydroxyls. However, the dependence of complex formation on p H has not always been investigated;'&I2 in some cases the p H values are not even reported, which makes it impossible to assess the composition of the complexes. This experimental study consists of three parts: (1) a study of the stoichiometry and equilibrium constants of the complexes 'This work is part of the Ph.D. thesis of I.P. at the Royal Institute of Technology.
0002-7863/87/1509-0380$01.50/0
formed in aqueous solution between Cu2+ and uridine, (2) the preparation of single crystals of a copper-uridine complex and the determination of its structure by means of X-ray diffraction, and (3) a correlation of the solution chemical data and the structure data by using small-angle X-ray scattering data (SAXS) from solutions of Cu2+ and uridine.
Experimental Section Chemicals. A stock solution of sodium perchlorate was prepared from sodium carbonate (Merck) and perchloric acid (Merck). Its salt content was determined by weighing samples dried at 110 O C . The copper(1I) perchlorate stock solution was prepared from copper(I1) oxide (Merck) and perchloric acid (Merck). The solution was analyzed for copper by electrodeposition. The H+ excess of these perchlorate stock solutions was determined potentiometrically by means of Gran13 plots. Sodium hydroxide solutions were prepared by diluting 50% NaOH (EKA) with boiled bidistilled water. Uridine (Serva) was dried to 110 O C . The solid copper-uridine compound used for the single-crystal diffraction measurements was prepared as described previo~sly.'~Anal. Calcd for C u ( ~ r i d ) ( c l O ~ ) ~62s(H20)6: ~ ~ ~ N a , c, 20.98; H, 4.69; N. 5.44; Cu, 12.33: C1, 4.30; Na, 7.25. Found: C, 20.51; H, 4.39; N, 5.72; Cu, 12.22: CI, 4.55; Na, 7.30. (1) (a) Laboratoire de Chimie de Coordination du CNRS. (b) Royal Institute of Technology. (c) University of Goteborg. (d) Chalmers University
of Technology. (2) (a) Eichhorn, G. L. Inorganic Biochemistry; Eichhorn, G. L., Ed.; Elsevier: Amsterdam, 1973; Vol. 2, pp 1210-1243. (b) Berger, N. A,; Tarien, E.; Eichhorn, G. L. Nature (London), New B i d . 1972, 239, 237. (3) Mosset, A.; Bonnet, J.4.; Galy, J. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1977, B33, 2639. (4) (a) Marzilli, L. G.; Kistenmacher, T. J. Acc. Chem. Res. 1977, 10, 146. (b) Gellert, R. W.; Bau, R. Met. Ions B i d . Syst. 1979, 8, 1. (5) (a) Phillips, R. Chem. Reu. 1966, 66, 501. (b) Izatt, R. M.; Christensen, J. J.; Rytting, J. H. Chem. Reu. 1971, 71, 439. (c) Martin, R. B. Acc. Chem. Res. 1985, 18, 32. (d) Marzilli, L. G.; de Castro, B.; Caradonna, J. P.; Steward, R. C.; Van Vuuren, C. P. J . A m . Chem. SOC.1980, 102, 916. (e) Martin, R. B.; Mariam, Y. H. Met. Ions Biol. Syst. 1979, 8, 57. ( f ) Pezzano, H.; Podo, F. Chem. Rec. 1980, 80, 365. (6) Fiskin, M.; Beer, M. Biochemistry 1965, 4, 1289. (7) Reinert, H.; Weiss, R. 2.Physiol. Chem. 1969, 350, 1321. (8) Chao, Y . - Y .H.; Kearns, D.R. J . A m . Chem. Sac. 1977, 99, 6425. (9) Chalilpoyil, P.; Marzilli, L. G . Inorg. Chem. 1979, 18, 2328. (10) Berger, N . A,; Eichhorn, G . L. Biochemistry 1971, 10, 1857. (11) (a) Kotowycz, G. Can. J . Chem. 1974, 52, 924. (b) Tu, A. T.; Friederich, C. G. Biochemistry 1968, 7, 4367. (c) Maskos, K. Acta Biochim. Pol. 1979, 26, 249. (12) Sovago, I.; Martin, R. B. Inorg. Chim. Acta 1980, 46, 91. (13),(a) Gran, G. Analyst (London) 1952, 77, 661. (b) Rossotti, F. J . C.; Rossotti, H. J . Chem. Educ. 1965, 42, 375. (c) Serjeant, E. P. Potentiometry and Potentiometric Titrations; Wiley: New York, 1984; p 225. (14) Bonnet, J.-J.;Jeannin, Y.: Mosset, A. C. R. Seances Acad. Sci.,Ser. C 1975, 280C, 827.
0 1987 American Chemical Society
J . Am. Chem. Soc., Vol. 109, No. 2, 1987
Unusual Copper-Uridine Octamer Table I. Crystallographic Data and Conditions for Data Collection
and Refinement Physical and Crystallographic Data formula: Cu8(urid)8Na8~5NaC104.48H20 molecular weight: 4098.7 crystal system: cubic space group: P432 a = 23.716 (5) 8, V = 13 339 A3 perp= 1.55 g cm+ pcalcd= 1.53 g cm-, absorption factor: p(hCuKa) = 30.5 cm-’ 1R 0.8 morphology: cube 0.25 X 0.25 X 0.25 mm ;-5
Data Collection temperature: 22 OC radiation: copper XCu K a = 1.54051 8, monochromator: oriented graphite crystal crystal-detector distance: 208 mm detector window: height” = 4 mm;width” = 4.00 + 0.35 tan 8 take-off angle: 5O scan mode: 8-28 maximum Bragg angle: 30° scan angle: A8 = AB, B tan 8; A8,“ = 1.6’; B” = 0.14 values determining the scan speed: SIGPRE” = 0.30; SIGMAa = 0.01; VPRE“ = 10’ mm-’ in 8; TMAX“ = 200 s controls: intensity reflections: 800; 033: 303 periodicity: 7200 s
+
Conditions for Refinement reflections for refinement of cell dimensions: 25 recorded reflections: 2167 independent reflections: 2009 utilized reflections: 1007 with I ? l.Su(r) reliability factors: “These parameters have been described in ref 3. Single-Crystal X-ray Data Collection and Structure Determination.
The data were collected at 22 OC on a CAD4 Enraf-Nonius automatic single-crystal diffractometer, by means of the 8-28 scan technique, using Cu K a radiation monochromatized with graphite, following the scheme in Table I and as described el~ewhere.~ During the course of the data collection, three standard reflections (Table I) were measured every 2 h, and the deviations of these reflections from their averages were all within counting statistics. Examination of a single crystal of the copper-uridine complex by means of precession technique showed that this compound belongs to the cubic system. The cubic crystalline system is very unusual with this class of compounds, and a rather complicated structure was expected. The observed Laue symmetry group m3m and the absence of systematic extinctions led to three possible space groups: P432, P43m, and Pm3m. Cell constants and corresponding standard deviations, listed in Table I, were derived from least-squares refinement of the settings of 25 automatically centered reflections. Based on a unit-cell volume of 13 339 A’, the calculated density (1.53 g cm-’) is in good agreement with the density of 1.55 g cm-, measured by flotation in a carbon tetrachloride and 1,1,2,2-tetrabromoethane medium. The shape of the selected crystal was a cube with bounding faces (loo), (OlO), and (001). The dimensions were 0.25 mm for the three edges. Among the 2009 unique reflections, 1007 have intensities I t 1.50(I), and only these reflections were included in subsequent refinements. The structure was solved by Patterson and Fourier methods and refined by full-matrix least-squares technique^.'^ The minimized quantity is Xw(lF,J - lFC1)*,where lFol and lFcl are the observed and calculated structure amplitudes and where the weights, w,are taken as 4F2/$(F,Z). The agreement indexes are defined in Table I. Values of the atomic scattering factors and the anomalous terms for the copper, sodium, and (15) In addition to local programs for the CII IRIS 80 computer, local modifications of the following programs were employed: Zalkin’s Fourier program, Ibers and Doedens’ NUCLS program, Busing, Martin, and Levy’s ORFFE program, and Johnson’s ORTEP program.
381
chlorine atoms were taken from Cromer and Waber’s tables.16 The statistical tests, performed on normalized structure factors, indicate that the space group is probably noncentrosymmetric. From the Patterson function, it was possible to deduce the copper atom position. The chosen space group is P432. From subsequent Fourier series, the ligand uridine and an octahedral Na(H,O)6 group were localized. Refinement of this partial solution gives reliability indexes R = 0.22 and R, = 0.30. At this point of the structural solution, two facts have to be pointed out: (a) the copper-uridine complex forms an octamer and the molecular species has the formula C ~ ~ ( u r i d ) ~ N a ( H and ~ O(b) ) ~in the huge cubic cell, these entities occupy the oxygen positions of a well-known simple solid-state structural type, the perovskite ABO,. This last remark allowed us to obtain the final structural solution by dividing the structure in two spatial volumes: B, an octahedron having C ~ ~ ( u r i d ) ~ N a ( at H its ~ ocorners, )~ and A, a cuboctahedron containing perchlorate anions, sodium cations, and water molecules. Partial Fourier synthesis was carried out on both the A and B volumes. In this way, we found one perchlorate anion in the B volume. The other atoms (Na atoms, water molecules, and perchlorate anions) are disordered in the A volume. We have not been able to refine all their positions. The disorder accounts for the poor quality of our diffraction data. Refinement of the positional parameters led to the final reliability index R = 0.13. We kept isotropic temperature factors because of the small number of reflections. Atomic parameters are listed in Table 11. The main interatomic distances and angles are given in Table 111. emf Titrations. Copper-uridine solutions with [H+] i= IO-” were titrated with perchloric acid to [H+] In order to test the equilibrium condition of the measurements, one experiment was performed by starting from a copper-uridine solution with [H’] = lo-* which was titrated with sodium hydroxide. Variations of activity coefficients and liquid junction potentials were minimized by keeping a constant ionic medium [C104-] = 3 M through the titrations. The emf of both a glass electrode (Metrohm EA107T) and a copper(I1) ion-selective electrode (Orion 94-29A) was measured against a double-junction Ag/AgCI reference electrode by using a digital voltmeter (Fluke 8810A) and an operational amplifier (Analog Devices 309K) in a minicomputer-controlled measuring system,” which also performed the additions from an automatic buret (Metrohm Dosimat). The titration vessel and reference electrode were immersed in a thermostated oil bath at 25.00 i 0.03 OC. The titrated solution was stirred with a magnetic bar, and nitrogen was used to prevent CO, gas contamination. The concentration range studied was 0.004-0.050 M for uridine and 0.0002-0.045 M for CU”. Both electrodes were calibrated by Gran’) plots, which also gave a check on the value of initial total proton excess. Unfortunately, the copper-selective electrode used did not function well in the pH region studied. This was true both for the commercial electrodes used and for the copper amalgam electrodes tested. The precision achieved in the emf measurements was 3 mV, corresponding to log [Cu2+] = 0.1. We do not know the reason for this low precision, as at lower pH one can easily attain a precision of 0.2 mV. One reason for the observed behavior might be adsorption of ligand at the electrode surface. Phenomena of this type are known to affect the behavior of the electrodes, * Small-Angle X-ray Scattering Measurements. The SAXS data were recorded with the monitor type camera developed by Kratky et aI.,l9 which was set up in a room thermostated at 21 O C . The sample container was a thin-walled Deybe-Sherrer glass capillary with an inner diameter of 0.96 mm. Monochromatization of the copper radiation was made with a nickel filter and a pulse-height discriminator together with a proportional counter. The absolute intensity was obtained by using the Lupolen method.20 The composition of the solutions used in the SAXS study are given in Table VI. For each sample, the background scattering of a solution containing a uridine concentration equal to [H3urid],o,- [Cu2+],,,,,and the same ionic composition, was measured and substracted from the scattering data. ;-5
’
(16) Cromer, D. T.; Waber, J. T. International Tables f o r X-ray Crystallography; Kynoch: Birmingham, England, 1974; Vol. IV, pp 72-98. (17) Cachet, J.; Wallin, T.; Zeising, J. TRITA-UUK-2013; Department of Inorganic Chemistry, Royal Institute of Technology, S-100 44 Stockholm,
Sweden, 1980. (18) (a) Brabec, V.; Christian, S. D.; Dryhurst, G. Eiophys. Chem. 1978, 7, 253. (b) Heijne, G. J. M.; van der Linden, W. E. Anal. Chim. Acta 1978, 96, 13. (19) Kratky, 0.;Leopold, H.; Seidler, H.-P. Z . Angew. Physik 1971, 31, 49. (20) Kratky, 0.;Pilz, I.; Schmitz, P. J. J . Colloid Interface Sci. 1966, 21, 24.
382 J. Am. Chem. SOC.,Vol. 109, No. 2, 1987
Gary et al.
Table 11. Final Atomic Parameters with esd’s Enclosed in Parentheses‘ atom X Y z 8 24 k 1 cu 0.1403 (3) 0.4512 (3) 0.1954 (3) 2.8 (1) 1 O(2) 0.0255 (13) 0.4244 (12) 24 k 0.1574.( 14) 3.1 (1) 1 o(4) 0.1264 (14) 24 k 0.4556 (12) 0.3227 (14) 3.0 (1) 24 k 1 O( 1 9 -0.0842 (9) 0.3330 (10) 0.2397 (9) 2.7 ( I ) 24 k 1 O(2’) -0.1341 (8) 0.4616 (8) 0.1935 (9) 2.9 (2) 24 k 1 O(3‘) -0.1855 (9) 0.3700 (9) 0.1634 (9) 3.2 (1) 24 k 1 O(5‘) -0.1268 (16) 0.2402 (15) 0.2896 (21) 4.9 (2) 1 N(1) -0.0229 (13) 0.4139 (10) 0.2423 (15) 24 k 2.9 (1) 1 N(3) 0.0717 (9) 0.4378 (7) 0.2427 (10) 2.4 ( I ) 24 k 0.4211 (9) 1 C(2) 0.0231 (10) 0.21 11 (1 1) 3.0 (1) 24 k 1 C(4) 0.0791 (11) 0.4363 (9) 0.3015 (12) 3.1 (2) 24 k 1 C(5) 0.0271 (11) 0.4325 (10) 0.3327 (13) 2.9 (1) 24 k 0.4170 (10) 0.3017 (12) 1 C(6) -0.0186 (10) 2.5 (1) 24 k 24 k 1 C(1‘) -0.0719 (9) 0.3868 (9) 0.2182 (10) 2.7 (2) 24 k 1 C(2‘) -0.1217 (11) 0.4202 (10) 0.2372 (13) 3.0 (1) 24 k 1 C(3’) -0.1703 (10) 0.3778 (9) 0.2221 (12) 3.1 (1) 24 k 1 (24‘) -0.1426 (12) 0.3207 (9) 0.2314 (13) 3.2 (1) 24 k 1 C(5‘) -0.1557 (14) 0.2881 (12) 0.2813 (15) 3.8 (2) 6‘ e 4 W l ) 0.4741 (8) 0 0 7.1 ( 1 ) 0.0686 ( I O ) 0.0686 (12) 10.7 (3) 12 j 2 W(1) ‘12 4 W(2) 0.3731 (11) 0 0 10.7 (3) 6 e 24 k 1 Na(2) 0.4351 (5) 0.1404 (5) 0.3951 (6) 6.3 (1) 6 f 4 W(3) II2 ‘I2 8.5 (2) 0.1402 (9) 0.2017 (1 1) 0.3199 (15) 1 W(4) 0.4021 (13) 9.1 (4) 24 k 1 W(5) 0.4451 (12) 0.2411 (10) 0.4152 (15) 9.0 (4) 24 k 3 C 42 C1( 1) ‘I2 ‘12 5.1 (1) 0 24‘ k 1 O(11) 0.4654 (35) 0.0351 (22) 0.4673 (29) 10.1 (4) 12 I 2 CU2) 0.1708 ( 6 ) 0 0.1708 (6) 5.8 ( I ) k 1 O(2 1) 0.1798 (15) -0.0012 (4) 0.1115 (18) 11.2 (3) 24 24 k 1 O(22) 0.1949 (14) 0.0511 (14) 0.1950 (18) 11 .o ( 3 ) “The first column indicates the number of positions, the Wyckoff notation, and the point symmetry. The positions of Na(3), W(6), W(7), and W(8) have not been refined. The sodium ion (position i) and two water molecules, W(6) and W(7) (position k), are located in the A zeolitic cage (cf. Figure 5). The W(8) water molecule (position k) is situated in the tunnel B. ‘50% occupancy. ~
~~
~
~~
~
~
~~~~
~~~
The analysis of the SAXS data followed conventional procedures involving correction for drift in the primary intensity and desmearing,2’ as well as model calculations using a weighted least-squares method.22 The radius of gyration, R,, and forward scattering, I(O), were obtained from Guinier plots.
Results A. Description of the Crystal Structure. The structure has two interesting aspects: (1) molecular entities CU8Uridg, containing Na(H,O), ions, surrounded by perchlorate anions and (2) a three-dimensional framework based on sodium atoms, giving rise to the formation of tunnels where the copper complexes are allocated, and cages containing perchlorate anions, sodium cations, and water molecules. I. Molecular Species Cu,urid8. The uridine ligand is threecoordinated toward three copper atoms: by the nitrogen atom N(3) of the pyrimidine base and by the oxygen atoms O(3’) and O(2’) of the ribose, this last oxygen atom bridging two copper atoms (Figure 1). The metal atoms are in a square-planar configuration and associated in dinuclear units in the CU8urid8 octamer species. The Cu-Cu distance within the dinuclear units is rather short (2.96
A).
The coordination of the nucleoside through the N(3) atom of the base has now been properly established in several structures of complexes containing transition metals and pyrimidine bases. However, to our knowledge, the double coordination of the ribose in the crystalline state is extremely unusual, although there is an example in an osmium bispyridine ester of a d e r ~ o s i n e . ~Such ~ a coordination for the copper-uridine system has been suggested8 from a ESR study in Me,SO solutions. However, in this study, a structure was proposed where the pyrimidine base does not participate in the coordination. The coordination of the pyrimidine base found in our study leads to distortions in the ligand geometry. The precision of the structure (21) Glatter, 0. J . Appl. Crystallogr. 1974, 7, 147. (22) Sjoberg, B. J . Appl. Crystallogr. 1978, 11, 7 3 . (23) Cohn, J. F.; Kim, J. J.; Suddath, F. L.; Blattmann, P.; Rich, A. J . Am. Chem. S o t . 1974, 96. 7152.
Figure 1. Coordination around the copper atom
refinement does not allow a precise comparison of bond lengths and angles; nevertheless, it appears interesting to discuss the values of the torsion angles (Table IV). The deviations from planarity (Table V) of the pyrimidine ring are larger than those in the free ligandz4 but smaller than in dihydr~uridine,’~ for example. The nucleoside conformation about the glycosidic bond N(l)-C( 1’) is anti, as in the free ligand and all the uridine analogues except 4 - t h i o ~ r i d i n e . The ~ ~ value of the corresponding torsion angle xCNis in the 0-72’ interval defined by Sundaralingam for the pyrimidine nucleoside^.^^ The puckering of the ribose ring is C(2’) endo, as shown by both the 0 angles (Table IV) and the positions of C(2’) and C(5’) (24) (a) Green, E. A,; Rosenstein, R. D.; Shiono, R.; Abraham, D. J.; Trus, B. L.; Marsh, R. E. Acta Crystallogr., Sect. B : Struct. Crystallogr. Cryst. Chem. 1975, 831, 102. (b) Green, E. A,; Rosenstein, R. D.; Shiono, R.; Abraham, D. J.; Trus, B. L.; Marsh, R. E. Acta Crystallogr., Sect. E : Struct. Crystallogr. Cryst. Chem. 1975, 8 3 1 , 1221. (25) Suck, D.; Saenger, W.; Zechmeister, K. Acta Crystallogr., Sect. B : Struct. Crystallogr. Cryst. Chem. 1972, B28, 596. (26) Saenger, W.; Scheit, K. H. J . Mol. E o [ . 1970, 50, 153. (27) Sundaralingam, M. Biopolymers 1969, 7, 821.
J . Am. Chem. SOC.,Vol. 109, No. 2, 1987 383
Unusual Copper-Uridine Octamer Table 111. Main Interatomic Distances and Angles“ Distances (8,) 1.94 C( 3’)-C( 2’) Cu-0(2’) 2.00 C(2’)-0(2’) Cu-N(3) 2.01 C(2’)-C( 1’) C~-0(3’) 2.07 C(4’)-C(5’) CU-O( 3’) C(5’)-O(5’) 1.33 Na( 1)-W( 1) N(1 )-C(2) 1.43 Na( 1)-W(2) ~(2)-~(3) 1.41 Na(2)-W(4) N(3)-C(4) 1.31 Na(2)-W(5) C(4)-0(4) 1.44 Na(2)-W(3) C(4)-C(5) 1.36 Na(2)-O(4) C(5)-C(6) 1.41 Na(2)-O(4) C(6)-N(l) Na(2)-O(11) I .45 N( 1)-C( 1’) 1.41 C1(1)-0( 11) C( 1’)-O( 1’) 1.43 C1(2)-0(21) O( 1’)-C(4’) 1.52 CI (2)-O( 22) C(4’)-C( 3’) 1.45 C( 3’)-O( 3’) Angles (:deg) 81.8 N( 1)-C(2)-0(2) 0(2’)-C~-0(3’) N( 1)-C(2)-N(3) 84.5 0(2’)-Cu-0(2’) 97.7 C(2)-N( 3)-C( 4) N(3)-Cu-O( 3’) 96.5 N(3)-C(4)-0(4) N( 3)-C~-0(2’) N( 3)-C(4)-C(5) C(4)-C( 5)-C(6) O( 1 l)-CI( 1)-O( 1 1) 108 107 C(5)-C(6)-N( 1) O(I l)-Cl(l)-0(11) C(6)-N( 1)-C( 2) 113 o(ll)-cl(l)-o(ll) C(6)-N( 1)-C( 1’) O(2 1)-C1(2)-0(2 1) 107 N( 1)-C( 1’)-C(2’) 110 O(2 l)-CI(2)-0(22) C( 1’)-C(2’)-0(2’) O(2 1 )-C1(2)-0(22) 108 112 C( 1’)-C(2’)-C( 3’) 0(21)-C1(2)-0(22) C(2’)-C(3’)-0(3’) 86 C(2’)-C( 3’)-c (4’) W ( 1 )-Na( 1)- W( 1) C( 3’)-C( 4’)-C( 5’) W(1)-Na( 1)-W(2) 105 C (3’)-C(4’)-0( 1’) 67 W(4)-Na(Z)-W(5) C(4’)-O( 1’)-C( 1’) W(4)-Na(2)-0(4) 97 O( 1’)-C( l’)-C(2’) 86 W(4)-Na(2)-0(4) 78 C(4’)-C ( 59-0 (5’) W(5)-Na(Z)-W(3) 96 W( 5)-Na(Z)-0(4) 97
~
1.57 1.46 1.49 1.45 1.34 2.38 2.40 2.43 2.45 2.92 3.04 3.13 3.11 1.40 1.42 1.45 127 114 128 119 I14 115 127 119 118 106 108 100 119 103 120 105 102 118
Table IV. Torsion Angles (deg)” Cu complex C(6)-N(l)-C(l’)-O(l’) 51 600 0(5’)-C(5’)-C(4’)-0( 1’) 60 175 @oc 0(5’)-C(S’)-C(4’)-C(3’) -42 8, C(4’)-0(l’)-C(l’)-C(2’) 44.5 8, O(l’)-C(1’)-C(2’)-C(3’) o3 C( l’)-C(Z’)-C(3’)-C(4’) -33 f14 C(2’)-C( 3’)-C( 4’)-O( 1’) 10 19.5 Bo c(3’)-c(4’)-0( 1’)-C( 1’) “The values of torsion angles for the free uridine are values.
%@I
Figure 2. The molecular species [C~~(urid)~Na(H~O),]’-.
110
“Average esd’s are 0.02 8, for Cu-0 and Cu-N bonds, 0.10 8, for other bonds, 0.6O for A-Cu-B angles, and 1 . 5 O for other bond angles.
XCN
Table VI. Composition of the Aqueous Solutions Used for the SAXS Measurements (mol/L) sample 1 2 3 4 0.0730 0.051 0.05 1 1 [Cu2+I101 0.0307 0.100 0.100 0.100 [H3urid],,, 0.100 00 0.20 0.22 0.20 [c104-1 0.10 0.30 0.32 “1 0.30 z11.3 z11.2 z11.6 211.1 -log IH’1
uridine 21.3 -75.3 42.7 7.0 -29.6 40.0 -37.0 19.0 the average
Table V. Least-Square Planeso
U1) N(1) -0.001 C(1’) 0.058 C(2) 0.063 O(1’) -0.099 O(2) -0.047 C(4‘) 0.110 N(3) -0.018 C(3‘) -0.073 -0.643 C(4) 0.1 11 c ( 2’) -0.830 o(4) -0.047 c ( 5’) C(5) -0.065 (76) 0.006 “Plane 1: N(1), C(2), 0 ( 2 ) , N(3). C(4), 0(4), C(5), C(6); 0.237X0.971Y + 0.3892 + 9.433 = 0. Plane 2: C(l’), O(l’), C(4’). C(3’), 0.043X- 0.252Y- 0.9612 + 7.597 = 0.
vs. the best mean plane of the ribose (Table V). Moreover, the two torsion angles @&o and 40-c show a gauche-trans conformation about the C(4’)-C(5’) bond. Thus, the ribose conformation
Figure 3. Stereoscopic view of a Cu,(urid),Na(H,O), entity. in the present complex is drastically different from the one observed in the free ligand24(where the puckering of the ribose ring is C(3’) endo and the conformation about the C(4’)-C(5’) bond is gauche-gauche) but very similar to the conformation of dihydrouridine.2s In the Cu8urid8 octamer, the eight metal atoms occupy the summits of a distorted Archimedian square antiprism (diagonal of the square fan: 11.4 A). The eight interconnected uridine ligands form a kind of toric structure (Figure 2 and 3). The external and internal diameters of the torus, calculated with the O(4) and O ( 2 ) oxygen atoms of the pyrimidine rings, are 16.4 and 7.6 A, respectively. There is an extensive overlap of the pyrimidine rings; the angle between two rings is about 6 O , and the mean distance is 3.67 A, the shortest one being 3.45 A (C(5)-C(5)). This overlap is much greater than the usual one defined by Bugg’s study.28 11. Molecular Species Na(H20)6. This octahedron of water molecules around a sodium atom occupies the center of the octamer Cu8urid8 (Figure 2). The O(2)-W( 1) distance (2.95 A) indicates the existence of hydrogen bonds 0-Ha-0 and that this N a ( H 2 0 ) 6 entity is thus locked by the O(2) atoms of the pyrimidine rings in the center of the Cu8urid8 toric structure. There are two positions, within this octahedron, statistically occupied a t 50% by the N a ( 1 ) cation. The coordination around the N a ( 1) atom is best described as a square pyramid, with the N a atom being displaced along the 4-fold axis from the square base toward the apex by 0.61 A. The sixth water molecule is then very weakly coordinated to the sodium cation. Such a coordination polyhedron for a sodium atom is not very common, although it (28) Bugg, C. E.; Thomas, J. M.; Sundaralingam, M.; Rao, S. T. Biopolymers 1971, IO, 175.
384
Gary et al.
J . Am. Chem. SOC.,Vol. 109, No. 2, 1987
Figure 4. The AB03 perovskite structure.
is found in synthetic or natural sodium silicates.29 111. Three-Dimensional Framework. The environment of the C1( 1) atom of the C104(1) group is a cube of oxygen atoms, the eight summits being half occupied. The ClO,( 1) group is thus statistically distributed between two positions and is centered on the faces of the cubic perovskite cell. The C104(2) groups occupy the summits of a cuboctahedron, centered on the cell origin (B cage of ABO,). The Na(2) atoms are at the center of a distorted square pyramid formed by O(4) oxygen atoms of the pyrimidine bases and water molecules. A sixth oxygen atom, which belongs to the C104(1) group, completes this polyhedron to give a distorted octahedron. We have used structure elements from two mineral structures, the ABO, pervoskite and the zeolites, to describe the structure. In the perovskite model (Figure 4), which has been of great help to solve the structure, the oxygen sites a r e occupied by the C ~ ~ ( u r i d ) ~ N a ( Hmolecular ~ O ) ~ species and the B sites by the formal association of 12 ClO, groups and one water molecule (W(8)). The A sites are filled by a very intricate sodium-water network. The Na(2) atoms also occupy the summits of a distorted polyhedron called snub-cube. Inside this polyhedron, we have a third N a + and two water molecules (W(6) and W(7)). These atoms are distributed on two succesive snub-cubes and a cuboctahedron, reminiscent of a nest of Russian dolls. However, the disorder makes it impossible to refine the positions, and we are not able to give a full description of the filling of the A sites. Nevertheless, this description requires a drastic idealization of the structure, especially for the A sites. The description in terms of the zeolite model is more satisfactory. The Na(2) snub-cubes are connected to each other in a zeolitic framework, quite similar to the framework of the synthetic Linde A zeolite (Figure 5 ) . The network of truncated octahedra connected by cubes (Linde A) is replaced by a network of snub-cubes connected by distorted Archimedian antiprisms in the present structure. Two kinds of cavities are thus defined: cages A, inside the snub-cubes, and tunnels B, parallel to the 4-fold axis. The mean diameter of these tunnels (about 16 A) is much larger than those observed in synthetic or natural zeolites (for example, 11.8 A30 in faujasite), which have similar cell dimensions ( a = 24.6 A). The connections between the snub-cubes, across the Archimedian antiprisms, are made by the ClO,(l) groups. In the tunnels B, the C ~ ~ ( u r i d ) ~ N a ( H molecular ~ o ) ~ species alternate with the ClO,(2) cuboctahedra. Thus, each molecular entity is surrounded by four ClO,( 1) groups at the corners of the equatorial plane of an octahedron, the apices of which are the ClO,(2) cuboctahedra. The position of the copper-uridine octamer, in the tunnel, is shown in Figure 6. This very intricate architecture, with a high symmetry, reminds one of the unusual copper(1)-copper(I1) complex of D-penicill(29) (a) Jost, K.-H.; Hilmer, W. Acta Crystallogr., Sect. E : Struct. Crystallogr. Cryst. Chem. 1966, 21, 583. (b) Perrault, G.; Boncher, C.; Vicat, J.; Cannillo, E.; Rossi, G. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1973, B29, 1432. (30) Barrer, R. M. Molecular Sieves; Meier, W. M., Uytterhoeven, J. B., Eds.; Advances in Chemistry 121; American Chemical Society: Washington, DC, 1973; pp 1-28.
Figure 5. Comparison between the sodium network and the Linde A zeolite structure.
Figure 6. Position of the copper-uridine complex in the tunnels
amine.31 This coordination compound crystallizes in the cubic system, and the structure consists of mixed valence clusters Cu,,L,, with a chloride ion locked at the center. These c~'8Cu"6L&15molecular species are surrounded by an extremely complicated network of thallium ions and water molecules (5 Tl+ and 55 H 2 0 per cluster unit). This thallium-water network is completely disordered as in our case. B. Complex Formation in Aqueous Solutions. Two different experimental methods have been used, small-angle X-ray scattering (SAXS) and potentiometric titrations. The equilibria will be described according to the expression qCu2+
+ sH3urid + H3,Cu,urid,2q-P + pHt
where H,urid stands for uridine. The equilibrium constants are (31) Birker, P. J. M. W. L.; Freeman, H. C. J . Chem. Sot., Chem. Commun. 1976, 312.
J . Am. Chem. SOC.,Vol. 109, No. 2, 1987
Unusual Copper-Uridine Octamer P
Bo 0 0
3.
Bo
__
Co
mM
4.2 4.1
11.73
A
7.3 15.0 4 5 . 0 Titration Bock 50.0
logtCu~'l-31ogIH'l+1og~H~uridl 7 Q 2.
CD
0.18
385
mM
4.7
._1 4 . 2 4.2 0.73 4.1
2.
0.10
7.3 15.0 45.0 50.0 Back Tltratlon
1.
1.0
1 predominate in the solutions. The absorption spectra of the [Cu2+] q[Cu,uridqe] = q[Cu,urid,9-]. Thus, we have plotted in Figure 8 the function -log [Cu2+]- 3 log [H+] + log [H3urid] dark-blue solutions, which show a band with a maximum near vs. log [Cu2+],,,, and these data can be used to test different 670 nm, do not change either with [H'] or with the concentration ratio of Cu(I1) to uridine, suggesting that only one copper complex models. q values between 4 and 8 give a satisfactory fit of the experimental data. The corresponding equilibrium constants are dominates in the solutions. I. emf Titrations. The acidity constant of uridine, which is log p12,4,4 = -68.4 and log &4,8,8 = -133.8. needed in later calculations, had to be determined first in cop11. SAXS Measurements. Table VI presents the chemical per-free solutions and was found to be log plol = log [H2urid-] composition of the samples studied; an inert ionic medium of + log [H+] - log [H3urid] = -9.66 0.01 (at 25 OC, [C104-] [C104-] = 0.2 mol/L was selected. The effect of the ionic medium = 3 M). was checked by studying one sample (No. 4, Table VI) without The results of the glass electrode measurements are shown in perchlorate, which was found to have a higher interparticle Figure 7, in the form p = ([H+] - [OH-] - [H+],,,)/[H3urid],,,, scattering effect than the rest of the samples. The effect of the Le., average p value per uridine vs. log [H+]. The continuous interparticle scattering was eliminated by extrapolating the curves were calculated assuming only one complex in the solutions I(m)/[Cu*+],,, values for samples 1, 2, and 3 to [Cu2+],,, = 0. with s = q and p / q = 3, Le., the values found in the solid phase After desmearing2' the extrapolated curve, we obtain a radius Cu8(urid),Na,.5NaC10,.48H20. From the glass electrode of gyration equal to 6.95 A. From the forward scattering I(O), measurements it is not possible to ascertain the value of q for the we calculate a molecular weight equal to 3.1 X l o 3 using the aqueous complex. Data of this type may be obtained from equation given by K r a t k ~and ~ ~a partial specific volume equal measurements of the free Cu2+ concentration, but the precision to 0.537 cm3/g. of our ion-selective electrode measurements is not enough to The extrapolated scattering curve is presented in Figure 9, where establish a unique chemical model for the Cu*+-uridine system. a comparison with some theoretical models is also given. Three However, from the precise log [H'] data we know that the reaction different theoretical models were tested: model a , a homogeneous taking place is spherical particle; model b, the atomic coordinates for the Cuuridine octamer ring found in the single crystal; and model c, the qCu2+ + qH3urid =+ Cu,urid,q3qHf atomic coordinates for the Cu-uridine octamer with the Na(H20), molecular unit in its center. From the equilibrium condition, we obtain log [Cu2+]- 3 log [H+] - s log
I
1-
+
~-
i
+
+
+
S
+
*
+
+ log [H+rid] = (1 / q ) log [Cu,urid,e] - (1 / q ) log pms,and from the mass balance condition for copper, we obtain [Cu2+],,, =
(32) Kratky, 0. Frog. Biophys. Mol. Biol. 1963, 13, 107.
I
386
J. Am. Chem. SOC.1987, 109, 386-390
A weighted least-squares method22was used to fit the models to the extrapolated scattering curve. Theoretical scattering curves for models b and c were calculated by using the Debye formula.33 A small, constant background was also fitted to the experimental data to compensate for the difference in ionic medium of the scattering curve compared with the background curve. Of the three models, the sphere model a gave the best fit, with a fitting index U (cf. eq 5 in ref 18) equal to 0.71% and a sphere radius equal to 8.63 A. The Uvalues obtained for models b and c were 3.57% and 1.40%, respectively. Thus we can conclude that the best fit with scattering data is obtained for a single sphere model, but an acceptable fit is also obtained for the octamer ring, with the Na(H,O), unit in the center. A sphere of radius ual to 8.63 8, corresponds to a radius of gyration equal to 6.68 , which is in agreement with the value, 6.95 A, obtained from Guinier plots after desmearing. From the forward scattering Z(0) obtained with the sphere model a, a molecular weight of 3.0 X IO3 is calculated, which is also in good agreement with the value 3.1 X l o 3 obtained from Guinier plots after desmearing. It may seem surprising that such a good fit is obtained from such a simplified model. However, similar observations have been made before.34 C . Discussion. The structures of the copper(I1)-uridine complex and its surrounding zeolitic framework are highly unusual. From the single-crystal structure determination it was not clear if the geometry of the complex is due to the unusual three-dimensional framework or if it is the geometry of the complex that determines the existence of the zeolitic framework. However, it seems chemically reasonable to expect such a structure for the copper(I1)-uridine complex. In fact, the coordination of metal ions by the 2’ and 3’ hydroxy oxygens of the ribose group has been proposed several times,7s8J2123and stacking of nucleosides in aqueous solutions is a well-known that favors the overlap of pyrimidine rings. Furthermore, several examples may be found of complexes with head-to-tail arrangements of purine or pyrimidine type ligand^.^^.^^
Furthermore, the S A X S data can be adequately explained both with a sphere of radius 8.63 A and with the molecular species [Cu8(urid)8Na(H20)6]7-,whose molecular weight, 2569, agrees fairly well with the value, 3.1 X lo3, obtained from the S A X S data for the particles in solution. It is also interesting that the introduction of the N a ( H 2 0 ) , unit into the center of the octamer ring gives a more than 6-fold decrease in the error square sum, U.The slightly better fit for the solid sphere of radius 8.63 A, as compared with the octamer ring, and the small difference in molecular weights might reflect either a certain degree of flexibility in the ring structure or, rather, the association of further Na+ ions to compensate for the charge of the complex. Thus, it should be concluded that the dominating Cu( 11)-uridine complex in solution has either the same structure as the octamer ring found in the single crystal or a similar one and that, therefore, the peculiar zeolitic framework found in the crystal is due to the geometric properties of the complex. This type of large polynuclear metal-nucleoside complex has not been reported previously, possibly due to the fact that they are difficult to detect in aqueous solution unless SAXS experiments are performed. Furthermore, single-crystal structure determinations of this kind of highly symmetrical and disordered solids are not a routine procedure. Although the situation in the tunnels can be regarded as a model of zeolitic sorption because of the relatively large “windows”, in fact, it results from a clathration process, since the liberation of the molecular species would induce a host-lattice breakdown. A related phenomena is the template synthesis of some zeolites with quaternary alkylammonium cations.37
(33) Debye, P. Ann. Phys. 1915, 46, 809. (34) Sjoberg, B.; Osterberg, R. Monatsh. Chem. 1982, 113, 915. (35) (a) Ts’o, P. 0.Basic Principles in Nuclei Acid Chemistry; Academic: New York, 1974; pp 537-584. (b) Borazan, H. N. J . Pharm. Sci. 1973,62, 1982. (c) Gallego, E.; Peral, F.; Morcillo, J. Anal. Quim. 1981, 77, 118. (36) (a) Faggiani, R.; Lippert, B.; Lock, C. J. L.; Speranzini, R. A. J . Am. Chem. SOC.1981, 103, 1 1 11. (b) Faggiani, R.; Lock, C. J. L.; Pollock, R. J.; Rosenberg, B.; Turner, G. Inorg. Chem. 1981, 20,804. (c) Berger, N. A.; Eichhorn, G. L. Biochemistry 1971, 10, 1847.
(37) (a) Breck, D. W. Zeolite Molecular Sieues; Wiley: New York, 1974; pp 304-312. (b) Flanigen, E.M. Molecular Sieoes; Meier, W. M., Uytterhoeven, J . B., Eds.; Advances in Chemistry 121; American Chemical Society: Washington, DC, 1973; pp 119-139. (c) Daniles, R. H.; Kerr, G. T.; Rollmann, L. D. J . A m . Chem. SOC.1978,100, 3097. (d) Wilson, S.T.; Lok, B. M.; Messina, C. A,; Cannan, T. R.; Flanigen, E. M. Intrazeolite Chemistry; Stucky, G. D., Dwyer, F. G., Eds.; ACS Symposium Series 218; American Chemical Society: Washington, DC, 1983; pp 79-106.
1
The results from the emf titrations show unequivocally that the following complex formation reactions take place in aqueous solutions: nCuZ+
+ nH3urid
Cu,urid,”-
+ 3nH’
~~~~
Some Aspects of the Reactivity of the First Stable Germaphosphene Jean Escudie, Claude Couret, Mbolatiana Andrianarison, and Jacques Satge* Contribution from the Laboratoire de Chimie des OrganominZraux, U.A. du C N R S no 477 UniversitZ Paul Sabatier, 31062 Toulouse Cedex, France. Received M a y 5, 1986
Abstract: Germaphosphene 1, the first stable compound with a germanium-phosphorus double bond, is very reactive toward compounds with active hydrogens producing secondary phosphines with a high regiospecificity and toward halogens to give adducts 11 and halogermylphosphines 12. Reaction with carbon tetrachloride allows, after addition followed by @-elimination, the quantitative synthesis of phosphaalkene 16. 1 reacts easily with nucleophilic compounds such as lithio compounds and Grignard reagents and probably via a radical process with dimethyl disulfide. The germanium-phosphorus double bond of 1 is reduced by lithium aluminum hydride and by the complex BH,.SMe, with formation of the germylphosphine 22; in the last case, a competitive addition of borane to the double bond was also observed.
Compounds of groups 14 and 15 in low coordination state are of particular interest a t the present time.’ Until recently me-
0002-7863/87/1509-0386$01.50/0
tallaphosphenes (>M=P-, M = Si, Ge, Sn) were postulated only as reactive intermediates and characterized by trapping reactions. 0 1987 American Chemical Society
